#!/usr/bin/perl

open(FILE, "$ARGV[0]")|| die "Csn not open $ARGV[0]!!\n";
open(OUT, ">>$ARGV[1]")|| die "Can not open $ARGV[1]!!\n";

$/= "\n#";
$tmp = $ARGV[0].".tmp";

while(<FILE>){
	chomp($_); 
	my @line = split("\n", $_);
	open(TMP, ">>$tmp");
	for my $i(0..$#line){
	#	if ($i == 1 || $i ==2 || $i == 4 ||$i ==5 || $i == 10 || $i ==11 || $i == 13 || $i == 14){
		if($line[$i] =~ /^>.*\s+(\S+)$/){
			print TMP ">$1\n$line[$i+1]\n\n";
		}
	#	}
	}
	system ("mafft-linsi --thread 3 --clustalout $tmp >$tmp.out");
	`rm $tmp`;
	open(TEMP, "$tmp.out");
	$/ = "\ng";
	undef($hum); undef($cow); undef($dog); undef($mouse); undef($rat); undef($star);
	while(my $k=<TEMP>){
		chomp($k);
		my @org = split("\n",$k);
		for my $i(0..$#org){
			if($org[$i] =~ /^rape\s+(\S+)$/){$hum .=$1; print "$1\n";}
	 		elsif($org[$i] =~ /^arabidopsis\s+(\S+)$/){$cow .=$1; print "$1\n";}
			elsif($org[$i] =~ /^moss\s+(\S+)$/){$dog .=$1; print "$1\n";}
			elsif($org[$i] =~ /^populus\s+(\S+)$/){$mouse .=$1; $str_ln=1; print "$1\n";}
			elsif($org[$i] =~ /^rice\s+(\S+)$/){$rat .=$1; $str_ln=1; print "$1\n";}
			elsif($str_ln>0){
				#if($org[$i] =~ /\*/){
					$str_ln=0; 
					$star .= substr($org[$i], 16, (length($org[$i])-16));
					#print "$star\n";
			}
		}
	}
	print OUT "> $line[0]\n$hum\n$cow\n$dog\n$mouse\n$rat\n$star\n\n";
	$/ = "\n#";
	`rm $tmp.out`;
}